.png)
The clinical reality few admit early enough
.png)
Every immuno-oncology (IO) program claims to be biomarker-driven, yet most fail to generate decision-grade pharmacodynamic (PD) data once real patient samples enter the equation.
If your PD assay can’t perform under clinical conditions, variable sample quality, multi-site logistics, and tight timelines. It won’t influence dose, escalation, or expansion.
Most biomarker failures happen before Phase II. They fail the moment they encounter real patient samples. Among 386 published Phase 1 IO trials, 100 (26%) reported no PD biomarker assessments at all; of the remainder, blood-based PD was reported in 270 studies and tissue-based PD in 94, while imaging-PD appeared in 12. Only 7-9% of PD readouts correlated with clinical activity, and just 8% were ever cited in subsequent trials⁹.
The 10 clinical reality checks outlined here distinguish biomarker programs that generate actionable insights from those that lose their signal before it can matter.
10 Clinical Reality Checks
1. Assay sensitivity and dynamic range
If the assay signal is too close to background levels, real biological changes can look inactive. In many protein biomarker tests, the lab standards used for calibration do not fully match the proteins found in patients.
As a result, tests using these artificial standards may not reflect how the assay performs with real clinical samples. To confirm true sensitivity, the method should be checked for consistency across different dilutions and verified directly on patient samples4,9.
2. Performance in real-world patient samples
Assays optimized on healthy donor samples or preclinical models often struggle when used on real patient material. Blood from patients can be hemolyzed, and tumor tissue can be partially necrotic or preserved in ways that change protein or cell quality. These factors can alter results even when the assay itself is technically sound5,2.
For that reason, performance should always be confirmed using actual patient samples to show that the assay works under real clinical conditions (Fountzilas et al., 2023; Bahassi, 2019). (Fountzilas et al., 2023; Bahassi, 2019).
3. Detectability of PD biomarkers in the patient population
For pharmacodynamic biomarkers, the key question is not only who has the target at baseline but how consistently that target shows measurable change after treatment.
If only a small proportion of patients, roughly 5 - 15%, show a detectable PD response, the study will lack enough data to link that signal to clinical outcomes or to optimize dose decisions. Blood sampling makes it possible to capture repeated measurements and detect more frequent pharmacodynamic shifts across patients, while tissue PD remains harder to collect and analyze in real time10,6.
4. Reproducibility across sites and operators
Getting consistent results in one laboratory does not guarantee that the assay will perform the same way across multiple trial sites.
Small differences in equipment, training, or sample handling can change PD readouts, especially when measurements depend on subtle biological shifts rather than high-concentration signals. To confirm reproducibility, studies should document precision, accuracy, and the working range of the assay in its intended clinical setting.
It is also important to recognize that proteins or cells in patient samples may behave differently from the reference materials used for calibration, which can lead to variability between sites if not properly controlled2,7.
5. Sample collection and stability
Many pharmacodynamic biomarkers fail not because of problems in the lab, but because the samples degrade before analysis. Temperature changes, shipping delays, or repeated freeze–thaw cycles can weaken or erase true biological signals.
Selecting the correct tube type, limiting processing time, and confirming stability through stress or freeze–thaw tests are essential to keep PD assay data reliable 9,10,4.
6. Fit-for-purpose analytical validation
Regulators now expect validation strategies that match how the biomarker will be used in the study. Traditional methods developed for pharmacokinetic assays only confirm analytical performance for drugs in plasma, not for biological signals that change with treatment.
Pharmacodynamic biomarkers, which measure natural molecules such as cytokines or immune-cell changes, require additional checks to show that the assay can detect real biological shifts in patient samples.
Guidance from the FDA (2018, 2025 Draft) and ICH M10 (2022) calls for this fit-for-purpose validation, including precision, accuracy, selectivity, range, and stability under actual clinical conditions 4,7.
7. Turnaround time and batch strategy
Pharmacodynamic data only have value if they are available when clinical teams need them. If results are delivered after dose-escalation meetings, safety reviews, or cohort expansion decisions, the data cannot guide the trial. Setting a realistic turnaround time, choosing when to batch samples, and coordinating with the data review cycle help ensure that PD findings influence dosing and development strategy2.
8. Longitudinal vs. baseline-only readouts
Single, baseline biomarker measurements rarely explain how a patient responds to immunotherapy. Pharmacodynamic biomarkers are most informative when they are tracked over time, showing how the immune system or tumor biology changes after each dose.
To make those time-based results meaningful, the assay must be validated to handle repeated sampling and to account for normal differences between and within patients. Understanding the expected variation within an individual and across the study population helps define what size of change represents a true biological effect rather than background noise 1,9.
9. Regulatory readiness and traceability
When pharmacodynamic data are used to support dose selection, mechanism of action, or response assessment, regulators now expect full transparency on how those data were generated. This includes a clear explanation of the scientific rationale for each biomarker, the validation status of the assay, and traceability of the data from sample collection through final analysis.
Incomplete documentation can slow regulatory review or cause biomarker-related endpoints to be questioned or removed from submissions. Maintaining clear records of assay version, validation reports, and sample handling builds confidence that PD results are both reliable and reproducible 6,7.
10. Endpoint alignment and interpretation
Pharmacodynamic biomarkers often fail not because the science is wrong, but because the assay is not linked to the trial’s clinical questions. When the PD readout does not clearly support decisions such as dose selection, safety review, or early signs of efficacy, the data end up classified as exploratory and are rarely used to guide the program.
In published Phase 1 immuno-oncology studies, measurable correlations between PD changes and patient outcomes were seen in only about 7 to 9 percent of cases, and just 8 percent of PD studies were cited in later trials. This shows that many PD assays produce interesting biological data but not information that drives dosing or development decisions5,9.
The trials that succeed tend to:
- Test assays on patient-derived samples, not just healthy donors or preclinical models
- Select biomarkers based on real prevalence and trial endpoints
- Ensure data turnaround aligns with dose-escalation and safety decisions
Roughly 70% of PD biomarker work relies on blood-based assays, yet few generate data that shape trial decisions. When validation, logistics, and timing align, PD biomarkers become decision tools, not just data points.
